Towards explicit description of ramification filtration in the 2-dimensional case (Q2483697)

For a positive integer \(n\) the field \(K\) is an \(n\)-dimensional local field if \(K\) is a complete discrete valuation field with residue field an \((n-1)\)-dimensional local field. A \(1\)-dimensional local field, or simply a local field, is a complete discrete valuation field with finite residue field and a \(2\)-dimensional local field is a complete discrete valuation field with residue field a local field. For a local field \(K\), let \(\Gamma:={\text{Gal}}(K_{\text{sep}}/ K)\) be the absolute Galois group of \(K\), where \(K_{\text{sep}}\) denotes a separable closure of \(K\). The classical ramification theory provides \(\Gamma\) with a decreasing filtration by the ramification subgroups \(\Gamma^{(v)}, v\geq 0\), \(\Gamma^ {(0)}\) being the inertia subgroup. The study of local fields is completely reduced to the study of \(\Gamma\) with this filtration. The author [Int. J. Math. 11, No. 2, 133--175 (2000; Zbl 1073.12501)] studied the \(1\)-dimensional local field of characteristic \(p\geq 3\) based on an explicit description of ramification filtration for maximal \(p\)-extensions with Galois groups of nilpotent class \(2\). The information obtained this way suffices to establish the property that the functor which associates \(K_ {\text{sep}}/K\) with \((\Gamma,\Gamma^{(v)})\) is fully faithful in this case and can be applied to the characteristic \(0\) case via the field of norms functor. In the paper under review, the author considers the \(2\)-dimensional case. He obtains the ramification filtration of \(\Gamma\) by the subgroups \(\Gamma^ {(v)}\) where \(v\) runs over the ordered set \(J=J_ 1\cup J_ 2\), ordered with the lexicographical order, and where \(J_ 1:= \{(v,c)\in {\mathbb Q}\times \{c\}\mid v\geq 0\}\), \(J_ 2:=\{\mu\in {\mathbb Q}^{2}\mid \mu\geq (0,0)\}\). The beginning of the filtration \(\{\Gamma^{(j)}\}_ {j\in J_ 1}\) comes from the classical \(1\)-dimensional ramification filtration of the group \(\Gamma_ c:={\text{Gal}}(K_ {c,{\text{sep}}}/K_ c)\) and its \(2\)-dimensional part \(\{\Gamma^{(\mu)}\}_{\mu\in J_ 2}\) gives a filtration of the group \(\widetilde{\Gamma}:= {\text{Gal}}(K_{\text{sep}}/K K_ {c,{\text{sep}}})\). Also, the beginning of the \(J_ 2\)-part comes from the classical ramification filtration of the group \({\text{Gal}}(K^{(1)}_ {\text{sep}}/K^ {(1)})\). In this paper it is given an explicit description of the image of the ramification \(\{\Gamma^{(j)}\}_{j\in J}\) in the maximal quotient of \(\Gamma\), which is a pro-\(p\)-group of nilpotent class \(2\) (Theorem 5.1). The method gives a generalization to the case of local fields of arbitrary dimension \(n\geq 2\). The result plays a clue role in the proof given by the author of the Grothendieck conjecture for higher dimensional local fields [Proc. Steklov Inst. Math. 241, 2--34 (2003); translation from Tr. Mat. Inst. Im. V. A. Steklova 241, 8--42 (2003; Zbl 1125.11355)]. The organization of the paper is as follows. In the first section is given the Artin-Schreier-Witt theory for the \(2\)-dimensional local field case. In Section 2, the author presents the \(2\)-dimensional ramification theory. Sections 3 and 4 deal with some auxiliary facts and filtrations. The last section presents the proof of the main result.